GEOID-Flood: Large-Scale Multi-Modal Benchmark for Flood Segmentation
links-ads · hf · 2026-08-04
Researchers introduce GEOID-Flood, a large-scale multi-modal benchmark for flood segmentation, addressing the lack of datasets combining bi-temporal SAR and optical imagery.
- Scale: Derived from Copernicus EMS, spanning 219 events across 65 countries with over 14,000 co-registered tiles (Sentinel-1, Sentinel-2, DEM).
- Findings: Foundation models offer a consistent but modest advantage; optical-SAR fusion with finetuning best resolves transient flooding; models trained on this dataset transfer better to unseen events.
More from Research
- Meta Open-Sources MHR: A High-Fidelity Parametric 3D Human Body Model — tom_doerr · 2026-08-04
- ARR August Cycle Receives 4,134 Submissions, Up 14% Year-over-Year — mdlhx · 2026-08-04
- Nature Publishes AI Biologist XunZi for Disease-Modifying Target Discovery — EricTopol · 2026-08-04
- Ai2 Launches OLMoEarth: Open-Source Planetary-Scale Geospatial Inference Platform — anselm · 2026-08-04
- Nature Study: AI Dermatology Diagnosis Amplifies Public Automation Bias — EricTopol · 2026-08-04
- Radical Co-founder on Training the Largest Genome Model to Write DNA — exnx · 2026-08-04